After pondering what would be suitable, I've submitted a Strawberry Perl talk for OSCON. It should be mostly an all-inclusive talk, covering a little history and lessons learned, current state (there should be another 2 releases before OSCON) and future direction, and provide an introduction to rolling your own Perl with Perl::Dist and building a product install for a Perl product (hopefully).
Other options included reruns of YAPC-esk talks like the PITA talk and Nothing Can Possibly Go Wrong (which I've already given in the US) but I get the impression that OSCON has a pretty full schedule, so I've just gone with the most current talk.
For YAPC::NA, where there's a bit more room for multiple talks, I will probably also add a talk I trialed at YAPC::AU called simply "::Tiny", and (assuming I can figure out an appropriate set of problems) I hope to present a sequel to one of my previous YAPC talks, which I plan to call "Nothing Can Possibly Go Wrong... Again".
This time around I plan to take a 3-10 year view of failure, and focus mostly on Economics and Biology (how to make your project violate the laws of nature, and lose).